Neural Networks and SVM-Based Classification of Leukocytes Using the Morphological Pattern Spectrum

Evaluation of thermal patterns and distribution applied to the study of diabetic foot

“…Erosion and dilation of an image are defined based on the interaction of a structuring element with the original image. The structuring element (λB) can have different forms, for example: square, circle, rectangle, and its size is defined by λ [16], [17]. Morphological operators for binary images are defined by:…”

Classification of chronic myeloid leukemia cell subtypes based on microscopic image analysis

“…In the past years, various computer-aided approaches based on image processing methods have been presented to classify five types of mature WBCs (Neutrophils, Eosinophils, Basophils, Lymphocytes and Monocytes) (Sarrafzadeh et al, 2014[ 29 ]; Habibzadeh et al, 2012[ 12 ]; Putzu et al, 2014[ 21 ]) six types of immature WBCs (CML cells; Myeloblast, Promyelocyte, Myelocyte, Metamyelocyte, Band and Neutrophil) (Theera-Umpon and Dhompongsa, 2007[ 33 ]; Ramirez-Cortes et al, 2010[ 22 ]; Theera-Umpon, 2005[ 32 ]) plasma cells in myeloma (Saeedizadeh et al, 2016[ 28 ]) and effective cells in diagnosis other leukemia (CLL (Alférez et al, 2014[ 3 ]), AML and ALL (Laosai and Chamnongthai, 2014[ 15 ]; MoradiAmin et al, 2016[ 19 ])).…”

“…Ramirez-Cortes et al (2010[ 22 ]) also conducted the segmentation of WBCs manually for classification of the six groups of CML cells, followed by the extraction of some texture and geometric features including pattern spectrum, normalized area, and nucleus to cytoplasm area ratio from 54 nuclei of WBCs. In this paper, the average of accuracy, 87.6 %, has been obtained by a Feed-forward Neural Networks (FFNN) classifier as the best result.…”

“…For example, some of these approaches need to calculate many image features that increase diagnostic time (Putzu et al, 2014[ 21 ]). Some of them (Ramirez-Cortes et al, 2010[ 22 ]; Saeedizadeh et al, 2016[ 28 ]) have only been evaluated on a few numbers of images or cells and sometimes the reported results of them did not have high performance (Theera-Umpon and Dhompongsa, 2007[ 33 ]). In addition, most of these approaches have just been evaluated and compared only by opinion of one expert (Sarrafzadeh et al, 2014[ 29 ]; Habibzadeh et al, 2012[ 12 ]; Putzu et al, 2014[ 21 ]; MoradiAmin et al, 2016[ 19 ]).…”

Automatic leukocytes classification by distance transform, moment invariant, morphological features, gray level co-occurrence matrices and SVM

“…J.M. Ramirez-Cortes et al [3] present the morphological operator pecstrum, or pattern spectrum, as a feature extractor of discriminating characteristics in microscopic leukocytes images for classification, their algorithm failed in classifying the eosinophil class, and the time cost is also a little large. Ko, B.C.…”